The Personal Fiduciary AI Opportunity to Drive Growth in UK Economy
Why I think UK is the best place to launch human-centric AI activity
Like many of in UK and beyond, I noted that UK.gov is ‘betting the farm’ on AI as the means to drive growth in the economy. I happen to think this a good thing. Clearly we must be quite precise as to what that means in practice, as almost every other ecomony will be doing the same. That means identifying which bits of the myriad of AI related opportunities we should back, and which not.
For me, this is a relatively easy choice and recommendation. In my view, UK.Gov should focus a big chunk of its AI effort on specialising in AI that specifically relates to personal data. You know, that bit that the rest of the world is studiously trying to avoid….
To bring that to life, the diagram below shows four segments, between then covering the whole spectrum of AI. For me that distinction is clear and sensible. Where AI is ingesting and/ or being applied to data for or about ‘things/ words/images etc ’, then there are no major barriers to doing so in terms of regulations. There are issues around copyright and similar, but the big regulations are largely avoided - hence the current wild west (or wild east) that we see now. That’s the bottom half of the chart, which will, I believe, be quickly commoditised; so I won’t dwell on it. There’s no shortage of other reading on the matter!!!
And then, the top half - AI and personal data. This is where I think there are massive opportunities; and that UK is almost uniquely placed to push hard on them.
So why is AI and personal data such a toxic, or at least scary territory anyway? That’s actually really straightforward, which I guess is why most large and small AI players have skirted around the area. The issue is that the data processing in and around AI that touches personal data renders data controller unable to meet their duties and obligations to the individual (data subject). Rights to transparency, access, rectification, portability and erasure are difficult enough at the best of times; but pretty much impossible once AI has been applied to the personal data - whether for training or model deployment. Those that have skirted around the issue have tended to do so at the edges via hidden updates to terms and conditions; these may ultimately prove costly. At the regulatory end, the most extreme and visible tend to be the Italian data protection regulator has issued bans on whole AI approaches.
That’s why AI that touches personal data requires pretty specialist skills, experience, and in many ways an attitudinal perspective that does not align well with current surveillance based approaches. And that’s why I think UK is well placed to take a leading position on what we’ll call Personal AI. My logic:
We have a bunch of ‘elder-statespeople’ who have been thinking along these lines for many years and have the scars to prove it. For example, Liz Brandt, Alan Mitchell, I and others set up what we called the Buyer Centric Commerce Forum way back in 2001 or so. Royal Mail looked at the issue of personal data empowerment at country scale back in 97-2000 and found the concept to be great and well received, but that in terms of the technology it was too early to do it well. That early work merged into Project VRM and in many similar initiatives worldwide; such as MyData Global; but the UK core has remained strong and active. The major, multi-year debate under the Blair Government (2006 or so) on whether UK should have a national identity scheme/ card or not also led to a significant education process for many in the population. I suspect we have hundreds of todays experts in privacy and data protection in UK who cut their teeth on that many pronged, very public identity scheme debate.
UK Gov understand the issue to at least a base level and has done for many years. The ‘MiData’ project was an early and important learning exercise around data portability from 2011-2013 or so. Again I suspect one main point to emerge would be that the technology (API’s) was also still not ready.
Then we have Open Banking, that massive investment in opening up what amounts to very sensitive personal financial data. Do I think we got all of that right; absolutely not (but we can fix that). But the whole process again means there are many people in UK who a) understand what it takes to design such schemes that are about moving data around; and then b) multiple million members of the population who are very familiar with the concept of making personal data available safely to third parties, and the permissions journey that implies and requires.
And the ‘Smart Data’ roadmap and plan has been building up now for several years, in many ways a successor to MiData with plans to drive data mobility (portability) across seven B2C sectors.
Then of course we have Brexit. Believe in it or not (I was not a fan); but it certainly gives UK the ability to be more flexible and fast moving around privacy and data protection than those locked into EU GDPR. But with more robustness than USA because we have UK GDPR as our start point. Yet we do get some ongoing benefits from historic EU presence, or at least market and geographic proximity to EU. Many of the big Internet Gatekeepers targeted by the EU Digital Markets Act for mandatory customer data portability have chosen to enable the same for UK market too. And those that don’t (and more) will get swept up in the same via the UK equivalent action from the Competition and Markets Authority.
So what might such a focus mean for UK in practice, and why the two separate segments within personal AI?
Let’s start with the easy one, AI ON People, by which I mean the large scale contributions of data that can be aggregated for a particular purpose which will now obviously wish to leverage AI. There are many examples of these typically ‘Data for Good’ scenarios. My personal favourite is UK Biobank. I’ve been a data donater since the start way back in 2006. Biobank has done a huge amount of good even before AI is pointed at the data sets; and could do vastly more if AI was engaged and the donation model updated. But here’s the rub….., i’ve stopped donating and granting new permissions; and from what I have heard i’m not the only one. Why is that? Very simple, the fairly regular requests to fill things in, or to grant consent for some organisation i’ve never heard of to access my data just does not get into my actual jobs to be done list. That’s resolvable, but not by doing more of the same. For that reason i’m see-ing this whole ‘data trust/ data/ AI for good as an area that UK can see as a real opportunity for growth. We have the knowledge and skills to do so if we focus them.
And then the BIG one. AI FOR People. I’ve written before about the opportunity around Personal Fiduciary AI Agents, so i’ll just point back to that detail here for those unfamiliar. But my point for this post is that if UK.Gov is looking for the areas in which UK has specific advantage, then Personal AI is probably the biggest - by miles. Most of the experts in the space know each other already; we are geographically relatively close, and we have hundreds if not thousands of organisations of all types who could support prototyping and then deployment. The Government aspect is also relatively closely defined in that the Department of Science, Innovation and Technology is clearly well placed to lead (as an extension/ branch of the Smart Data Roadmap).
I’m doing my bit with upcoming launches for MyKey and DataPal. MyKey is a fundamental building block that all in the human-centric model can leverage. DataPal is way out there in capability terms as a building block for Personal AI in that the richness, quality, depth and provenance of data individuals can bring to the AI in un-paralleled.
I’m not quite sure on how best to foster the wider collaboration; hence writing about the opportunity. Maybe a new MyData UK Hub as has been discussed; maybe some other routes. I’m open to suggestions. Hopefully the huge opportunity and the economic necessity in the country can drive that collaboration.